Clinical utility of custom-designed NGS panel testing in pediatric tumors

Abstract

Background: Somatic genetic testing is rapidly becoming the standard of care in many adult and pediatric cancers. Previously, the standard approach was single-gene or focused multigene testing, but many centers have moved towards broad-based next-generation sequencing (NGS) panels. Here, we report the laboratory validation and clinical utility of a large cohort of clinical NGS somatic sequencing results in diagnosis, prognosis, and treatment of a wide range of pediatric cancers.

Methods: Subjects were accrued retrospectively at a single pediatric quaternary-care hospital. Sequence analyses were performed on 367 pediatric cancer samples using custom-designed NGS panels over a 15-month period. Cases were profiled for mutations, copy number variations, and fusions identified through sequencing, and their clinical impact on diagnosis, prognosis, and therapy was assessed.

Results: NGS panel testing was incorporated meaningfully into clinical care in 88.7% of leukemia/lymphomas, 90.6% of central nervous system (CNS) tumors, and 62.6% of non-CNS solid tumors included in this cohort. A change in diagnosis as a result of testing occurred in 3.3% of cases. Additionally, 19.4% of all patients had variants requiring further evaluation for potential germline alteration.

Conclusions: Use of somatic NGS panel testing resulted in a significant impact on clinical care, including diagnosis, prognosis, and treatment planning in 78.7% of pediatric patients tested in our institution. Somatic NGS tumor testing should be implemented as part of the routine diagnostic workup of newly diagnosed and relapsed pediatric cancer patients.

Keywords: Molecular profiling; Pediatric cancer; Tumor sequencing.

Fig. 1

Copy number alterations observed in 389 pediatric tumors. The number of observed instances of copy number gains and losses is represented on the X axis vs. the chromosomal band on the Y axis. Tumor types are color-coded. CNS = central nervous system

Fig. 2

Fusions detected in 389 pediatric tumors. Circos plots showing 5′ (blue) and 3′ (red) fusion genes across three categories of tumors tested. Genes are listed adjacent to chromosomal number and location. Central lines connect fusion partners. Details of novel/rare fusions identified can be found in Additional file 1: Table S7. CNS = central nervous system

Fig. 3

Solid tumor (non-CNS) oncoprint of most common Tier 1 and 2 SNVs, idels, and fusions. Summary of most commonly encountered Tier 1 or 2 SNVs and fusion variants identified in > 1 solid tumor. TP53 was the most commonly encountered SNV, followed by KRAS, ALK, and BRCA2. EWSR1-FLI1 was the most common fusion gene, followed by FOXO1 fusions, present in a total of four tumors. Overall number of clinically significant variants per tumor is represented across the top, with eight tumors have > 1 alteration. The number of variants identified per gene is represented to the right. Age and tumor type are represented across the bottom. CNS = central nervous system, SNV = single nucleotide variant, indel = insertion/deletion

Fig. 4

CNS Tumor oncoprint of most common Tier 1 and 2 SNVs, idels, and fusions. Summary of most commonly encountered Tier 1 or 2 SNVs and fusion variants identified in > 1 CNS tumor. BRAF fusions and point mutations (V600E) were most common, followed by TP53 and H3F3A hotspot Lys28Met variants. Tumors without BRAF alterations were more likely to have > 1 variant identified per sample, which occurred in 22 tumors. Overall number of clinically significant variants per tumor is represented at the top. The number of variants identified per gene is represented to the right. Age and tumor type are represented across the bottom. CNS = central nervous system, SNV = single nucleotide variant, indel = insertion/deletion

Fig. 5

Leukemia/lymphoma oncoprint of most common Tier 1 and 2 SNVs, indels, and fusions. Summary of most commonly encountered Tier 1 or 2 SNVs and fusion variants identified in > 1 leukemia/lymphoma tumor. NRAS and KRAS hotspot variants were most common, followed by ETV6-RUNX1 fusions. Compared to solid tumors and CNS tumors, Leukemia and lymphomas were more likely to have multiple variants per tumor, represented across the top. The number of variants identified per gene is represented to the right. Age and tumor type are represented across the bottom. CNS = central nervous system, SNV = single nucleotide variant, indel = insertion/deletion

Fig. 6

Clinical impact of panel sequencing. Cases were considered positive for clinical impact (+) if they contained clinically significant Tier 1 or 2 variants. a. Overall, 289 cases (78.7%) had positive clinical impact for at least one type of category (diagnostic, prognostic, therapeutic, potential germline). The bar chart on the right shows the percentage of cases with clinically significant Tier 1 or 2 variants impacting each category. b. Summary of clinical impact color-coded by tumor type and impact category. Pie chart depicts number of patients in each slice. Hash marks indicate no clinical impact while solid color indicate positive clinical impact